Energy storage system charging and discharging strategy optimization method based on big data analysis

Abstract

The invention relates to the technical field of energy storage systems, in particular to an energy storage system charging and discharging strategy optimization method based on big data analysis. The technical problems are that the energy utilization rate and the charging and discharging efficiency of the energy storage system are low due to the fact that a traditional charging and discharging strategy of the energy storage system is difficult to adapt to a dynamically changing power grid environment and user requirements, and the normal operation and the service life of the system are affected due to the fact that key components of the energy storage system are prone to faults due to lack of effective fault prediction and monitoring means. According to the technical scheme, the energy storage system charging and discharging strategy optimization method based on big data analysis comprises the following steps: establishing a time sequence prediction model and a particle swarm optimization model based on a data analysis result; according to the method, the future state and demand of the energy storage system can be accurately predicted through big data analysis and a machine learning algorithm, so that the optimal charging and discharging strategy is formulated; the energy utilization rate of the energy storage system is improved, the charging and discharging efficiency is remarkably improved, and the operation cost is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of energy storage systems, and in particular to a method for optimizing charging and discharging strategies of energy storage systems based on big data analysis. Background Art

[0002] With the continuous growth of energy demand and the rapid development of renewable energy, the role of energy storage systems in power systems has become increasingly prominent. Energy storage systems can not only effectively balance the contradiction between supply and demand and improve the stability of power systems, but also provide emergency backup power in the event of power grid failures. However, how to efficiently and safely manage and operate energy storage systems, especially in the face of complex and changing power grid environments and user needs, has become a technical problem that needs to be solved urgently.

[0003] Traditional energy storage system charging and discharging strategies are often based on simple rules or experience, lacking real-time and intelligence, and are difficult to adapt to the dynamically changing grid environment and user needs. This leads to low energy utilization, low charging and discharging efficiency, and potential safety hazards in energy storage systems. In addition, due to the lack of effective fault prediction and monitoring methods, key components of energy storage systems are prone to failure, affecting the normal operation and life of the system.

[0004] Therefore, there is an urgent need for an energy storage system charging and discharging strategy optimization method based on big data analysis to improve the energy utilization, charging and discharging efficiency and safety of the energy storage system, reduce operating costs and extend system life. Summary of the invention

[0005] In order to overcome the problems of lack of intelligence in the charging and discharging strategy of the energy storage system, low energy utilization, low charging and discharging efficiency, and insufficient fault prediction in the existing technology.

[0006] The technical solution of the present invention is: a method for optimizing the charging and discharging strategy of an energy storage system based on big data analysis, comprising the following steps:

[0007] S1: Identify key data collection points in the energy storage system, deploy sensors and monitoring equipment, and integrate and transmit data to the data analysis platform;

[0008] S2: Preprocess the collected raw data, including cleaning data, converting data formats, and normalizing and standardizing data;

[0009] S3: extract key features from the data preprocessed in step S2, use machine learning algorithms to perform data analysis on the extracted key features, explore the patterns and trends in the data, and identify potential problems of the energy storage system; wherein the extracted key features include:

[0010] Statistical characteristics: including mean, variance, maximum, minimum, kurtosis, and skewness, which are used to capture the basic statistical characteristics of data and reflect the distribution and change trend of data;

[0011] Trend and seasonal characteristics: Time series trend and seasonal components are used to understand how data changes over time;

[0012] Frequency characteristics: The dominant frequency and energy distribution characteristics in the spectrum obtained by Fourier transform of time series data are used to reveal the periodic components and energy distribution in the data;

[0013] Time-frequency features: The key features of the time-frequency spectrum extracted using the short-time Fourier transform method are used to reflect the changes in data in the time and frequency domains;

[0014] Correlation features: The correlation between the features and the target variable is evaluated by the chi-square test method. The features with higher correlation are selected to identify factors that have a greater impact on the state or key parameters of the energy storage system;

[0015] S4: Based on the data analysis results in step S3, a time series prediction model and a particle swarm algorithm optimization model are established;

[0016] S5: According to the results of the time series prediction model and the particle swarm algorithm optimization model output, formulate a preliminary charging and discharging strategy, including the key parameters of charging time, discharging time, and charging and discharging power; the key parameters are determined by:

[0017] Charging time: Based on the battery capacity and charging power, as well as the prediction of battery charging demand by the time series prediction model, the appropriate charging time can be determined. If it is predicted that the battery usage will be large in the future, the charging time can be increased in advance to ensure that the battery has sufficient power.

[0018] Discharge time: The discharge time also depends on the battery capacity and discharge power, as well as the prediction of the battery discharge demand by the time series prediction model. By predicting the battery discharge demand, the discharge time can be arranged to avoid excessive discharge of the battery or a long period of low power.

[0019] Charging and discharging power: Based on the battery specifications, the capacity of the charging facilities, the discharge load requirements, and the prediction of the battery charging and discharging requirements by the time series prediction model, the charging and discharging power is optimized to ensure that the battery meets the load requirements during the charging and discharging process;

[0020] S6: Use the historical data and real-time data of the energy storage system operating parameters to simulate and verify the preliminary charging and discharging strategy formulated in step S5, evaluate the performance and effect of the preliminary charging and discharging strategy, and adjust and optimize the charging and discharging strategy based on the simulation verification results;

[0021] S7: Use big data technology to establish a fault prediction model. By deploying sensors and monitoring equipment on key components of the energy storage system, the real-time status data of the energy storage system is input into the fault prediction model to monitor and predict the key components of the energy storage system in real time. If the model predicts a fault risk, an early warning signal is sent to notify the operator to promptly discover and handle potential faults.

[0022] S8: Deploy sensors and monitoring equipment on key components of the energy storage system to monitor the energy storage system in real time, collect operating data and feed it back to the data analysis platform;

[0023] S9: Regularly evaluate and adjust the charging and discharging strategy.

[0024] As a preferred method, identify key data collection points in the energy storage system, deploy sensors and monitoring equipment, and integrate and transmit data to the data analysis platform; the specific steps are as follows:

[0025] S201: By deploying sensors and monitoring equipment on key components of the energy storage system, key parameters of the energy storage system operation are extracted. The key parameters include battery pack: voltage, current, temperature, state of charge, health status; PCS: input / output power, efficiency, working mode; BMS: battery cell voltage, current, temperature, balance status; control system: charge and discharge instructions, system status, fault alarm;

[0026] S202: Select relevant sensors and monitoring equipment based on the identified key parameters and deploy them at key locations of the energy storage system; the relevant sensors and monitoring equipment include:

[0027] Battery pack monitoring sensors: voltage sensors, current sensors, temperature sensors, state of charge (SOC) sensors, and state of health (SOH) monitoring devices;

[0028] PCS monitoring equipment: power sensors, efficiency monitors and operating mode indicators;

[0029] BMS monitoring equipment: battery cell monitoring module, balancing controller, overvoltage monitoring equipment, overcurrent monitoring equipment and overtemperature monitoring equipment;

[0030] Control system monitoring equipment: charge and discharge instruction recorder and system status monitor;

[0031] S203: Integrate the data collected by various sensors and monitoring devices to form a unified data set;

[0032] S204: Connect the integrated data to the data analysis platform using wireless transmission.

[0033] As a preferred method, clean the data, convert the data format, and perform data normalization and standardization. The specific steps are as follows:

[0034] S301: Remove invalid data: Check whether there is invalid data in the data set. For null values, delete the record and fill it with the average value and median.

[0035] S302: Remove duplicate data: Check whether there are duplicate records in the data set, and use a data deduplication algorithm based on a hash table to quickly identify and delete duplicate data;

[0036] S303: Convert the timestamps of different data into a readable time format, and convert the data type into a readable format according to the actual meaning of the data field;

[0037] S304: scaling the data to a range of 0 to 1 to eliminate the effects of different dimensions;

[0038] S305: Convert the data into a standard normal distribution form with a mean of 0 and a standard deviation of 1 to perform data standardization.

[0039] As a preferred method, key features are extracted from the preprocessed data, and machine learning algorithms are used to analyze the data, mine the patterns and trends in the data, and identify potential problems of the energy storage system. The specific steps are as follows:

[0040] S401: Calculate the statistics of the data, including mean, variance, maximum, minimum, kurtosis, and skewness, to capture the basic statistical characteristics of the data;

[0041] S402: Extract trend and seasonal components of time series using moving average method;

[0042] S403: Perform Fourier transform on the time series data to obtain a spectrum, analyze the frequency components, and extract the dominant frequency and energy distribution characteristics in the spectrum;

[0043] S404: Extract key features from the time-frequency spectrum using short-time Fourier transform method;

[0044] S405: Use the chi-square test method to evaluate the correlation between the features and the target variable, select the features with higher correlation, and remove irrelevant and redundant features;

[0045] S406: Using the extracted and selected features as input variables, and using the state and key parameters of the energy storage system as target variables;

[0046] S407: Divide the data set into a training set and a test set;

[0047] S408: Initialize the random forest model and set parameters including the number of trees, maximum depth, and minimum number of sample splits;

[0048] S409: Use the training set data to train the random forest model and construct multiple decision trees. Each tree randomly selects features for splitting during training.

[0049] S410: Use the trained random forest model to predict the test set data, calculate the accuracy, recall rate, and F1 score indicators of the prediction results, and evaluate the performance of the model;

[0050] S411: Analyze the importance of each feature in the random forest model, identify the features that have a greater impact on the state or key parameters of the energy storage system, and identify potential problems of the energy storage system based on the importance of the features.

[0051] As a preferred method, based on the data analysis results, a time series prediction model and a particle swarm algorithm optimization model are established; the specific steps are as follows:

[0052] S501: According to the data processed in step S3, a time series to be predicted is selected to build a time series prediction model;

[0053] S502: Construct the autoregressive part of the prediction model. The formula is as follows:

[0054] X t =φ1X t-1 +φ2X t-2 +...+φ p X t-p +∈ t ;

[0055] Among them, X t is the value of the time series at time t, φ i is the autoregressive coefficient, p is the autoregressive order, ∈ t It is white noise;

[0056] S503: Construct the moving average part of the prediction model. The formula is as follows:

[0057] ∈ t =θ1∈ t-1 +θ2∈ t-2 +...+θ q ∈ t-q ;

[0058] Among them, θ i is the moving average coefficient, q is the moving average order;

[0059] S504: combining the autoregressive and moving average parts to form an ARIMA (p, d, q) model, where d is the difference order;

[0060] S505: Use time series data to estimate the parameters of the ARIMA model, including autoregressive coefficients, moving average coefficients, and difference orders, and use the least squares method to optimize the parameters;

[0061] S506: According to the trained ARIMA model, the future time series data is predicted, and the accuracy of the prediction results is evaluated using the mean square error;

[0062] S507: randomly generate a group of particles, each particle represents a potential charging and discharging strategy solution, including key parameters of charging time, discharging time, and charging and discharging power;

[0063] S508: Initialize the speed and position of each particle, where the speed represents the moving direction and size of the particle in the search space, and the position represents the current solution of the particle;

[0064] S509: Use the time series prediction model to predict the state of the energy storage system in the future;

[0065] S510: Calculate the fitness value of each particle according to the prediction result;

[0066] S511: Update the speed and position of each particle according to the fitness value of the particle and the global optimal solution, where the speed update formula is as follows:

[0067]

[0068] in, is the velocity of particle i at the k+1th iteration, ω is the inertia weight, c1 and c2 are learning factors, r1 and r2 are random numbers, is the personal optimal solution of particle i, g k Make the global optimal solution, is the current position of particle i;

[0069] The position update formula is as follows:

[0070]

[0071] S512: Repeat S509-S511 until the upper limit of the number of iterations is reached or the fitness value meets the preset conditions, and output the global optimal solution.

[0072] As a preferred method, a preliminary charging and discharging strategy is formulated according to the results of the prediction model and the optimization model, including key parameters of charging time, discharging time, and charging and discharging power; the specific steps are as follows:

[0073] S601: Obtain the energy storage system status prediction for a period of time in the future from the time series prediction model, including but not limited to the state of charge, temperature, and power demand of the battery pack; identify the peak demand, valley period, and possible abnormal conditions that may occur in the future, and determine the feasible range and safety threshold of charging and discharging according to the characteristics and limitations of the energy storage system;

[0074] S602: Setting initial charging and discharging strategy parameters, including charging time, discharging time, and charging and discharging power;

[0075] S603: Optimizing the charge and discharge strategy parameters based on the predicted data using a particle swarm optimization algorithm;

[0076] S604: Continuously adjust the charging and discharging strategy parameters through iterative calculation until an optimal solution is found or a preset upper limit of iteration times is reached, and output the result;

[0077] S605: Formulate a detailed charging and discharging plan according to the optimized charging and discharging strategy parameters, including specific charging time, discharging time, and charging and discharging power.

[0078] As a preferred method, historical data and real-time data are used to simulate and verify the strategy, evaluate the performance and effect of the strategy, and adjust and optimize the strategy according to the simulation verification results; the specific steps are as follows:

[0079] S701: Build a simulation platform to simulate the operation of the energy storage system, including the charging and discharging behavior of the battery pack, the working status of the PCS, and the monitoring function of the BMS;

[0080] S702: Importing historical data and real-time data into the simulation platform;

[0081] S703: according to the established charging and discharging strategy, setting simulation parameters, including charging time, discharging time, and charging and discharging power;

[0082] S704: Setting the start and end time of the simulation, as well as the time step of the simulation;

[0083] S705: Starting the simulation platform to simulate the operation of the energy storage system according to the set parameters and strategies;

[0084] S706: Real-time recording of state variables of the energy storage system during the simulation process, including the state of charge, temperature, voltage, current, etc. of the battery pack, and the input and output power and efficiency of the PCS;

[0085] S707: Compare the simulation results with the expected goals, evaluate the performance of the charging and discharging strategy, and calculate key performance indicators, including the energy utilization rate, charging and discharging efficiency, temperature control effect, and power response speed of the energy storage system;

[0086] S708: Analyze abnormal situations or unexpected behaviors that occur during the simulation process and identify possible problem points;

[0087] S709: adjusting and optimizing the charge and discharge strategy according to the simulation evaluation results, including adjusting key parameters of the charge time, discharge time, and charge and discharge power;

[0088] S710: Repeat the simulation until the performance of the strategy reaches the expected target or cannot be further improved;

[0089] S711: Output the optimized charging and discharging strategy as executable instructions and schedules for use by the energy storage system controller.

[0090] As a preferred method, a fault prediction model is established using big data technology to monitor and predict key components of the energy storage system in real time, and to detect and handle potential faults in a timely manner. The specific steps are as follows:

[0091] S801: Use random forest model to build fault prediction model;

[0092] S802: Use the training set divided by historical data to train the model, use the trained random forest model to predict the test set data, and adjust the model parameters and feature selection according to the performance of the model on the test set to optimize the performance of the model;

[0093] S803: Deploy the trained fault prediction model to the monitoring platform of the energy storage system to monitor and predict key components in real time;

[0094] S804: Deploy sensors and monitoring equipment on key components of the energy storage system to collect real-time operating data, including voltage, current, temperature, pressure, vibration parameters, as well as working status and mode operation information;

[0095] S805: The collected data is transmitted to the data analysis platform in real time via wireless means;

[0096] S806: Input the real-time collected data into the random forest fault prediction model to perform real-time fault prediction. The model will predict whether there is a risk of failure of key components based on the changes and trends of the data;

[0097] S807: When the model predicts a failure risk, the system sends a warning signal to notify the operator.

[0098] Preferably, when constructing the random forest fault prediction model, the following steps are included:

[0099] S901: For each decision tree, randomly select a portion of features for splitting, and calculate the splitting point of each feature;

[0100] S902: Divide the data set into different subsets according to the splitting points, and recursively build a subtree on each subset to obtain a complete decision tree;

[0101] S903: Prediction stage, input new data into the random forest model, and each decision tree will give a prediction result;

[0102] S904: adopting a voting mechanism to select the prediction result of the majority decision tree as the final prediction result.

[0103] As a preferred method, the energy storage system is monitored in real time, operation data is collected and fed back to the data analysis platform, and the charging and discharging strategy is evaluated and adjusted regularly. The specific steps are as follows:

[0104] S1001: Sensors and monitoring equipment collect operating data of the energy storage system in real time;

[0105] S1002: Perform preliminary data cleaning at the data collection end to remove obviously erroneous or invalid data points, and transmit the pre-processed data to the data analysis platform in real time;

[0106] S1003: On the data analysis platform, the energy storage system operation data collected in real time is stored in a database;

[0107] S1004: Use data visualization tools to display the operating status and key parameters of the energy storage system in real time:

[0108] S1005: setting an evaluation cycle of a charging and discharging strategy according to the operating characteristics and actual needs of the energy storage system;

[0109] S1006: At the end of the evaluation cycle, collect and organize the operation data of the energy storage system during the cycle, including real-time collected data, historical data, and simulation data;

[0110] S1007: Analyze the implementation effect of the charging and discharging strategy during the cycle using big data analysis, including key performance indicators of the energy storage system, such as energy utilization rate, charging and discharging efficiency, temperature control effect, and power response speed;

[0111] S1008: Compare the system performance changes before and after the implementation of the strategy, evaluate the effectiveness of the strategy, and identify problems in the operation of the energy storage system based on the strategy effect analysis;

[0112] S1009: According to the problem identification result, the charging and discharging strategy is adjusted and optimized accordingly, including adjusting key parameters of charging time, discharging time, and charging and discharging power;

[0113] S1010: Use the simulation platform to verify the adjusted strategy and evaluate its performance and effect;

[0114] S1011: Update the verified optimized charging and discharging strategy to the control system of the energy storage system.

[0115] Beneficial effects of the present invention:

[0116] 1. Significantly improve the energy utilization rate and charging and discharging efficiency of the energy storage system: Through big data analysis and machine learning algorithms, the present invention can accurately predict the future state and demand of the energy storage system, thereby formulating the optimal charging and discharging strategy; it not only improves the energy utilization rate of the energy storage system, but also significantly improves the charging and discharging efficiency and reduces the operating cost;

[0117] 2. Effectively extend the service life of the energy storage system: The present invention uses big data technology to establish a fault prediction model, which can monitor and predict the fault conditions of key components of the energy storage system in real time, help to timely discover and deal with potential faults, avoid the expansion of faults and cause system damage, and thus effectively extend the service life of the energy storage system;

[0118] 3. Improve the safety and stability of the energy storage system: Through real-time monitoring and data analysis, the present invention can promptly discover abnormal conditions and potential safety hazards in the operation of the energy storage system. Combined with optimized charging and discharging strategies, the present invention can effectively avoid safety issues such as overcharging and over-discharging, and improve the overall safety and stability of the energy storage system. At the same time, through regular strategy evaluation and adjustment, the present invention can also adapt to the needs of the energy storage system that change with time and environment, and maintain its long-term and efficient operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0119] Figure 1 What is shown is a schematic flow chart of the steps of the energy storage system charging and discharging strategy optimization method based on big data analysis of the present invention;

[0120] Figure 2 What is presented is a schematic flow chart of the working steps of the particle swarm algorithm optimization model of the energy storage system charging and discharging strategy optimization method based on big data analysis of the present invention. DETAILED DESCRIPTION

[0121] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0122] See also Figure 1-2 The present invention provides an embodiment: a method for optimizing a charging and discharging strategy of an energy storage system based on big data analysis, comprising the following steps:

[0123] S1: Identify key data collection points in the energy storage system, deploy sensors and monitoring equipment, and integrate and transmit data to the data analysis platform;

[0124] S2: Preprocess the collected raw data, including cleaning data, converting data formats, and normalizing and standardizing data;

[0125] S3: extract key features from the data preprocessed in step S2, use machine learning algorithms to perform data analysis on the extracted key features, explore the patterns and trends in the data, and identify potential problems of the energy storage system; wherein the extracted key features include:

[0126] Statistical characteristics: including mean, variance, maximum, minimum, kurtosis, and skewness, which are used to capture the basic statistical characteristics of data and reflect the distribution and change trend of data;

[0127] Trend and seasonal characteristics: Time series trend and seasonal components are used to understand how data changes over time;

[0128] Frequency characteristics: The dominant frequency and energy distribution characteristics in the spectrum obtained by Fourier transform of time series data are used to reveal the periodic components and energy distribution in the data;

[0129] Time-frequency features: The key features of the time-frequency spectrum extracted using the short-time Fourier transform method are used to reflect the changes in data in the time and frequency domains;

[0130] Correlation features: The correlation between the features and the target variable is evaluated by the chi-square test method. The features with higher correlation are selected to identify factors that have a greater impact on the state or key parameters of the energy storage system;

[0131] S4: Based on the data analysis results in step S3, a time series prediction model and a particle swarm algorithm optimization model are established;

[0132] S5: According to the results of the time series prediction model and the particle swarm algorithm optimization model output, formulate a preliminary charging and discharging strategy, including the key parameters of charging time, discharging time, and charging and discharging power; the key parameters are determined by:

[0133] Charging time: Based on the battery capacity and charging power, as well as the prediction of battery charging demand by the time series prediction model, the appropriate charging time can be determined. If it is predicted that the battery usage will be large in the future, the charging time can be increased in advance to ensure that the battery has sufficient power.

[0134] Discharge time: The discharge time also depends on the battery capacity and discharge power, as well as the prediction of the battery discharge demand by the time series prediction model. By predicting the battery discharge demand, the discharge time can be arranged to avoid excessive discharge of the battery or a long period of low power.

[0135] Charging and discharging power: Based on the battery specifications, the capacity of the charging facilities, the discharge load requirements, and the prediction of the battery charging and discharging requirements by the time series prediction model, the charging and discharging power is optimized to ensure that the battery meets the load requirements during the charging and discharging process;

[0136] S6: Use the historical data and real-time data of the energy storage system operating parameters to simulate and verify the preliminary charging and discharging strategy formulated in step S5, evaluate the performance and effect of the preliminary charging and discharging strategy, and adjust and optimize the charging and discharging strategy based on the simulation verification results;

[0137] S7: Use big data technology to establish a fault prediction model. By deploying sensors and monitoring equipment on key components of the energy storage system, the real-time status data of the energy storage system is input into the fault prediction model to monitor and predict the key components of the energy storage system in real time. If the model predicts a fault risk, an early warning signal is sent to notify the operator to promptly discover and handle potential faults.

[0138] S8: Deploy sensors and monitoring equipment on key components of the energy storage system to monitor the energy storage system in real time, collect operating data and feed it back to the data analysis platform;

[0139] S9: Regularly evaluate and adjust the charging and discharging strategy.

[0140] As a preferred method, identify key data collection points in the energy storage system, deploy sensors and monitoring equipment, and integrate and transmit data to the data analysis platform; the specific steps are as follows:

[0141] S201: By deploying sensors and monitoring equipment on key components of the energy storage system, key parameters of the energy storage system operation are extracted. The key parameters include battery pack: voltage, current, temperature, state of charge, health status; PCS: input / output power, efficiency, working mode; BMS: battery cell voltage, current, temperature, balance status; control system: charge and discharge instructions, system status, fault alarm;

[0142] S202: Select relevant sensors and monitoring equipment based on the identified key parameters and deploy them at key locations of the energy storage system; the relevant sensors and monitoring equipment include:

[0143] Battery pack monitoring sensors: voltage sensors, current sensors, temperature sensors, state of charge (SOC) sensors, and state of health (SOH) monitoring devices;

[0144] PCS monitoring equipment: power sensors, efficiency monitors and operating mode indicators;

[0145] BMS monitoring equipment: battery cell monitoring module, balancing controller, overvoltage monitoring equipment, overcurrent monitoring equipment and overtemperature monitoring equipment;

[0146] Control system monitoring equipment: charge and discharge instruction recorder and system status monitor;

[0147] S203: Integrate the data collected by various sensors and monitoring devices to form a unified data set;

[0148] S204: Connect the integrated data to the data analysis platform using wireless transmission.

[0149] As a preferred method, clean the data, convert the data format, and perform data normalization and standardization. The specific steps are as follows:

[0150] S301: Remove invalid data: Check whether there is invalid data in the data set. For null values, delete the record and fill it with the average value and median.

[0151] S302: Remove duplicate data: Check whether there are duplicate records in the data set, and use a data deduplication algorithm based on a hash table to quickly identify and delete duplicate data;

[0152] S303: Convert the timestamps of different data into a readable time format, and convert the data type into a readable format according to the actual meaning of the data field;

[0153] S304: scaling the data to a range of 0 to 1 to eliminate the effects of different dimensions;

[0154] S305: Convert the data into a standard normal distribution form with a mean of 0 and a standard deviation of 1 to perform data standardization.

[0155] As a preferred method, key features are extracted from the preprocessed data, and machine learning algorithms are used to analyze the data, mine the patterns and trends in the data, and identify potential problems of the energy storage system. The specific steps are as follows:

[0156] S401: Calculate the statistics of the data, including mean, variance, maximum, minimum, kurtosis, and skewness, to capture the basic statistical characteristics of the data;

[0157] S402: Extract trend and seasonal components of time series using moving average method;

[0158] S403: Perform Fourier transform on the time series data to obtain a spectrum, analyze the frequency components, and extract the dominant frequency and energy distribution characteristics in the spectrum;

[0159] S404: Extract key features from the time-frequency spectrum using short-time Fourier transform method;

[0160] S405: Use the chi-square test method to evaluate the correlation between the features and the target variable, select the features with higher correlation, and remove irrelevant and redundant features;

[0161] S406: Using the extracted and selected features as input variables, and using the state and key parameters of the energy storage system as target variables;

[0162] S407: Divide the data set into a training set and a test set;

[0163] S408: Initialize the random forest model and set parameters including the number of trees, maximum depth, and minimum number of sample splits;

[0164] S409: Use the training set data to train the random forest model and construct multiple decision trees. Each tree randomly selects features for splitting during training.

[0165] S410: Use the trained random forest model to predict the test set data, calculate the accuracy, recall rate, and F1 score indicators of the prediction results, and evaluate the performance of the model;

[0166] S411: Analyze the importance of each feature in the random forest model, identify the features that have a greater impact on the state or key parameters of the energy storage system, and identify potential problems of the energy storage system based on the importance of the features.

[0167] As a preferred method, based on the data analysis results, a time series prediction model and a particle swarm algorithm optimization model are established; the specific steps are as follows:

[0168] S501: According to the data processed in step S3, a time series to be predicted is selected to build a time series prediction model;

[0169] S502: Construct the autoregressive part of the prediction model. The formula is as follows:

[0170] X t =φ1X t-1 +φ2X t-2 +...+φ p X t-p +∈ t ;

[0171] Among them, X t is the value of the time series at time t, φ i is the autoregressive coefficient, p is the autoregressive order, ∈t is white noise;

[0172] S503: Construct the moving average part of the prediction model. The formula is as follows:

[0173] ∈ t =θ1∈ t-1 +θ2∈ t-2+...+θ q ∈ t-q ;

[0174] Among them, θ i is the moving average coefficient, q is the moving average order;

[0175] S504: combining the autoregressive and moving average parts to form an ARIMA (p, d, q) model, where d is the difference order;

[0176] S505: Use time series data to estimate the parameters of the ARIMA model, including autoregressive coefficients, moving average coefficients, and difference orders, and use the least squares method to optimize the parameters;

[0177] S506: According to the trained ARIMA model, the future time series data is predicted, and the accuracy of the prediction results is evaluated using the mean square error;

[0178] S507: randomly generate a group of particles, each particle represents a potential charging and discharging strategy solution, including key parameters of charging time, discharging time, and charging and discharging power;

[0179] S508: Initialize the speed and position of each particle, where the speed represents the moving direction and size of the particle in the search space, and the position represents the current solution of the particle;

[0180] S509: Use the time series prediction model to predict the state of the energy storage system in the future;

[0181] S510: Calculate the fitness value of each particle according to the prediction result;

[0182] S511: Update the speed and position of each particle according to the fitness value of the particle and the global optimal solution, where the speed update formula is as follows:

[0183]

[0184] in, is the velocity of particle i at the k+1th iteration, ω is the inertia weight, c1 and c2 are learning factors, r1 and r2 are random numbers, is the personal optimal solution of particle i, g k Make the global optimal solution, is the current position of particle i;

[0185] The position update formula is as follows:

[0186]

[0187] S512: Repeat S509-S511 until the upper limit of the number of iterations is reached or the fitness value meets the preset conditions, and output the global optimal solution.

[0188] As a preferred method, a preliminary charging and discharging strategy is formulated according to the results of the prediction model and the optimization model, including key parameters of charging time, discharging time, and charging and discharging power; the specific steps are as follows:

[0189] S601: Obtain the energy storage system status prediction for a period of time in the future from the time series prediction model, including but not limited to the state of charge, temperature, and power demand of the battery pack; identify the peak demand, valley period, and possible abnormal conditions that may occur in the future, and determine the feasible range and safety threshold of charging and discharging according to the characteristics and limitations of the energy storage system;

[0190] S602: Setting initial charging and discharging strategy parameters, including charging time, discharging time, and charging and discharging power;

[0191] S603: Optimizing the charge and discharge strategy parameters based on the predicted data using a particle swarm optimization algorithm;

[0192] S604: Continuously adjust the charging and discharging strategy parameters through iterative calculation until an optimal solution is found or a preset upper limit of iteration times is reached, and output the result;

[0193] S605: Formulate a detailed charging and discharging plan according to the optimized charging and discharging strategy parameters, including specific charging time, discharging time, and charging and discharging power.

[0194] As a preferred method, historical data and real-time data are used to simulate and verify the strategy, evaluate the performance and effect of the strategy, and adjust and optimize the strategy according to the simulation verification results; the specific steps are as follows:

[0195] S701: Build a simulation platform to simulate the operation of the energy storage system, including the charging and discharging behavior of the battery pack, the working status of the PCS, and the monitoring function of the BMS;

[0196] S702: Importing historical data and real-time data into the simulation platform;

[0197] S703: according to the established charging and discharging strategy, setting simulation parameters, including charging time, discharging time, and charging and discharging power;

[0198] S704: Setting the start and end time of the simulation, as well as the time step of the simulation;

[0199] S705: Starting the simulation platform to simulate the operation of the energy storage system according to the set parameters and strategies;

[0200] S706: Real-time recording of state variables of the energy storage system during the simulation process, including the state of charge, temperature, voltage, current, etc. of the battery pack, and the input and output power and efficiency of the PCS;

[0201] S707: Compare the simulation results with the expected goals, evaluate the performance of the charging and discharging strategy, and calculate key performance indicators, including the energy utilization rate, charging and discharging efficiency, temperature control effect, and power response speed of the energy storage system;

[0202] S708: Analyze abnormal situations or unexpected behaviors that occur during the simulation process and identify possible problem points;

[0203] S709: adjusting and optimizing the charge and discharge strategy according to the simulation evaluation results, including adjusting key parameters of the charge time, discharge time, and charge and discharge power;

[0204] S710: Repeat the simulation until the performance of the strategy reaches the expected target or cannot be further improved;

[0205] S711: Output the optimized charging and discharging strategy as executable instructions and schedules for use by the energy storage system controller.

[0206] As a preferred method, a fault prediction model is established using big data technology to monitor and predict key components of the energy storage system in real time, and to detect and handle potential faults in a timely manner. The specific steps are as follows:

[0207] S801: Use random forest model to build fault prediction model;

[0208] S802: Use the training set divided by historical data to train the model, use the trained random forest model to predict the test set data, and adjust the model parameters and feature selection according to the performance of the model on the test set to optimize the performance of the model;

[0209] S803: Deploy the trained fault prediction model to the monitoring platform of the energy storage system to monitor and predict key components in real time;

[0210] S804: Deploy sensors and monitoring equipment on key components of the energy storage system to collect real-time operating data, including voltage, current, temperature, pressure, vibration parameters, as well as working status and mode operation information;

[0211] S805: The collected data is transmitted to the data analysis platform in real time via wireless means;

[0212] S806: Input the real-time collected data into the random forest fault prediction model to perform real-time fault prediction. The model will predict whether there is a risk of failure of key components based on the changes and trends of the data;

[0213] S807: When the model predicts a failure risk, the system sends a warning signal to notify the operator.

[0214] Preferably, when constructing the random forest fault prediction model, the following steps are included:

[0215] S901: For each decision tree, randomly select a portion of features for splitting, and calculate the splitting point of each feature;

[0216] S902: Divide the data set into different subsets according to the splitting points, and recursively build a subtree on each subset to obtain a complete decision tree;

[0217] S903: In the prediction phase, new data is input into the random forest model, and each decision tree will give a prediction result;

[0218] S904: adopting a voting mechanism to select the prediction result of the majority decision tree as the final prediction result.

[0219] As a preferred method, the energy storage system is monitored in real time, operation data is collected and fed back to the data analysis platform, and the charging and discharging strategy is evaluated and adjusted regularly. The specific steps are as follows:

[0220] S1001: Sensors and monitoring equipment collect operating data of the energy storage system in real time;

[0221] S1002: Perform preliminary data cleaning at the data collection end to remove obviously erroneous or invalid data points, and transmit the pre-processed data to the data analysis platform in real time;

[0222] S1003: On the data analysis platform, the energy storage system operation data collected in real time is stored in a database;

[0223] S1004: Use data visualization tools to display the operating status and key parameters of the energy storage system in real time:

[0224] S1005: setting an evaluation cycle of a charging and discharging strategy according to the operating characteristics and actual needs of the energy storage system;

[0225] S1006: At the end of the evaluation cycle, collect and organize the operation data of the energy storage system during the cycle, including real-time collected data, historical data, and simulation data;

[0226] S1007: Analyze the implementation effect of the charging and discharging strategy during the cycle using big data analysis, including key performance indicators of the energy storage system, such as energy utilization rate, charging and discharging efficiency, temperature control effect, and power response speed;

[0227] S1008: Compare the system performance changes before and after the implementation of the strategy, evaluate the effectiveness of the strategy, and identify problems in the operation of the energy storage system based on the strategy effect analysis;

[0228] S1009: According to the problem identification result, the charging and discharging strategy is adjusted and optimized accordingly, including adjusting key parameters of charging time, discharging time, and charging and discharging power;

[0229] S1010: Use the simulation platform to verify the adjusted strategy and evaluate its performance and effect;

[0230] S1011: Update the verified optimized charging and discharging strategy to the control system of the energy storage system.

[0231] Example 1

[0232] Optionally, assume that there is a set V containing battery pack voltage data, where V = [V1, V2, ..., Vn], and the steps for normalizing this data set are as follows:

[0233] First, find the minimum value V in the data set min and maximum value V max ;

[0234] Then, use the normalization formula: Where i represents the i-th element in the data set;

[0235] Through calculation, the normalized data set is as follows:

[0236]

[0237] Preferably, the above steps can be used to clean, convert, normalize and standardize the raw data, providing a high-quality data foundation for subsequent data analysis, modeling and optimization of charging and discharging strategies.

[0238] Example 2

[0239] Optionally, a data set D is set to contain n samples and m features, where each sample is a state record of an energy storage system, and the features include the voltage, current, and temperature of the battery pack;

[0240] When training the model, the steps are as follows:

[0241] A1: Initialize the random forest model, set the number of trees to T, the maximum depth to d, and the minimum number of sample splits to s.

[0242] A2: For each tree t (t=1, 2, ..., T):

[0243] a. Randomly extract a subset D_t from the data set D as training data;

[0244] b. Randomly select k features from m features as candidate split features (k<m);

[0245] c. According to the candidate splitting features, select the optimal splitting point for node splitting and build a decision tree;

[0246] d. Repeat steps b and c until the maximum depth d is reached or other stopping conditions are met;

[0247] A3: Obtain T decision trees to form a random forest model;

[0248] When making model predictions and evaluations, the steps are as follows:

[0249] B1: For each sample x in the test set, use each tree in the random forest model to make a prediction and get T prediction results;

[0250] B2: Use majority voting mechanism to determine the final prediction result y pred ;

[0251] B3: Calculate the accuracy, recall, and F1 score of the prediction results to evaluate the performance of the model;

[0252] When identifying potential issues, the steps are as follows:

[0253] C1: Analyze the importance of each feature in the random forest model, using indicators such as feature importance score or feature contribution;

[0254] C2: Based on the importance of features and combined with domain knowledge, potential problems of the energy storage system are identified, such as abnormal voltage fluctuations in the battery pack, which may indicate battery aging or improper charge and discharge control.

[0255] Preferably, through the above steps, it is possible to extract key features from the preprocessed data, use machine learning algorithms such as random forests to analyze the data, explore the patterns and trends in the data, and identify potential problems of the energy storage system.

[0256] The embodiments of the present invention are described in detail above with reference to the accompanying drawings, but the present invention is not limited to the above embodiments, and various changes can be made within the knowledge scope of those skilled in the art without departing from the spirit of the present invention.

Claims

1. A method for optimizing charging and discharging strategies of energy storage systems based on big data analysis, characterized in that: The following steps are included: S1: Identify key data collection points in the energy storage system, deploy sensors and monitoring equipment, and integrate and transmit data to the data analysis platform; S2: Preprocess the collected raw data, including cleaning data, converting data formats, and normalizing and standardizing data; S3: extract key features from the data preprocessed in step S2, use machine learning algorithms to perform data analysis on the extracted key features, explore the patterns and trends in the data, and identify potential problems of the energy storage system; wherein the extracted key features include: Statistical characteristics: including mean, variance, maximum, minimum, kurtosis, and skewness, which are used to capture the basic statistical characteristics of data and reflect the distribution and change trend of data; Trend and seasonal characteristics: Time series trend and seasonal components are used to understand how data changes over time; Frequency characteristics: The dominant frequency and energy distribution characteristics in the spectrum obtained by Fourier transform of time series data are used to reveal the periodic components and energy distribution in the data; Time-frequency features: The key features of the time-frequency spectrum extracted using the short-time Fourier transform method are used to reflect the changes in data in the time and frequency domains; Correlation features: The correlation between the features and the target variable is evaluated by the chi-square test method. The features with higher correlation are selected to identify factors that have a greater impact on the state or key parameters of the energy storage system; S4: Based on the data analysis results in step S3, a time series prediction model and a particle swarm algorithm optimization model are established; S5: According to the results of the time series prediction model and the particle swarm algorithm optimization model output, formulate a preliminary charging and discharging strategy, including the key parameters of charging time, discharging time, and charging and discharging power; the key parameters are determined by: Charging time: Based on the battery capacity and charging power, as well as the prediction of battery charging demand by the time series prediction model, the appropriate charging time can be determined. If it is predicted that the battery usage will be large in the future, the charging time can be increased in advance to ensure that the battery has sufficient power. Discharge time: The discharge time also depends on the battery capacity and discharge power, as well as the prediction of the battery discharge demand by the time series prediction model. By predicting the battery discharge demand, the discharge time can be arranged to avoid excessive discharge of the battery or a long period of low power. Charging and discharging power: Based on the battery specifications, the capacity of the charging facilities, the discharge load requirements, and the prediction of the battery charging and discharging requirements by the time series prediction model, the charging and discharging power is optimized to ensure that the battery meets the load requirements during the charging and discharging process; S6: Use the historical data and real-time data of the energy storage system operating parameters to simulate and verify the preliminary charging and discharging strategy formulated in step S5, evaluate the performance and effect of the preliminary charging and discharging strategy, and adjust and optimize the charging and discharging strategy based on the simulation verification results; S7: Use big data technology to establish a fault prediction model. By deploying sensors and monitoring equipment on key components of the energy storage system, the real-time status data of the energy storage system is input into the fault prediction model to monitor and predict the key components of the energy storage system in real time. If the model predicts a fault risk, an early warning signal is sent to notify the operator to promptly discover and handle potential faults. S8: Deploy sensors and monitoring equipment on key components of the energy storage system to monitor the energy storage system in real time, collect operating data and feed it back to the data analysis platform; S9: Regularly evaluate and adjust the charging and discharging strategy.

2. The method for optimizing the charging and discharging strategy of an energy storage system based on big data analysis according to claim 1, characterized in that: Identify key data collection points in the energy storage system, deploy sensors and monitoring equipment, and integrate and transmit data to the data analysis platform; the specific steps are as follows: S201: Deploy sensors and monitoring equipment on key components of the energy storage system to extract key parameters of the energy storage system operation, including battery pack: voltage, current, temperature, state of charge, and health status; PCS: input / output power, efficiency, operating mode; BMS: battery cell voltage, current, temperature, and balance status; control system: charge and discharge instructions, system status, and fault alarm; S202: Select relevant sensors and monitoring equipment based on the identified key parameters and deploy them at key locations of the energy storage system; the relevant sensors and monitoring equipment include: Battery pack monitoring sensors: voltage sensor, current sensor, temperature sensor, state of charge (S0C) sensor and health state (S0H) monitoring device; PCS monitoring equipment: power sensors, efficiency monitors and operating mode indicators; BMS monitoring equipment: battery cell monitoring module, balancing controller, overvoltage monitoring equipment, overcurrent monitoring equipment and overtemperature monitoring equipment; Control system monitoring equipment: charge and discharge instruction recorder and system status monitor; S203: Integrate the data collected by various sensors and monitoring devices to form a unified data set; S204: Connect the integrated data to the data analysis platform using wireless transmission.

3. The method for optimizing the charging and discharging strategy of an energy storage system based on big data analysis according to claim 2 is characterized in that: Clean the data, convert the data format, and perform data normalization and standardization. The specific steps are as follows: S301: Remove invalid data: Check whether there is invalid data in the data set. For null values, delete the record and fill it with the average value and median. S302: Remove duplicate data: Check whether there are duplicate records in the data set, and use a data deduplication algorithm based on a hash table to quickly identify and delete duplicate data; S303: Convert the timestamps of different data into a readable time format, and convert the data type into a readable format according to the actual meaning of the data field; S304: scaling the data to a range of 0 to 1 to eliminate the effects of different dimensions; S305: Convert the data into a standard normal distribution form with a mean of 0 and a standard deviation of 1 to perform data standardization.

4. The method for optimizing the charging and discharging strategy of an energy storage system based on big data analysis according to claim 3 is characterized in that: Extract key features from the preprocessed data, use machine learning algorithms to analyze the data, mine the patterns and trends in the data, and identify potential problems of the energy storage system; the specific steps are as follows: S401: Calculate the statistics of the data, including mean, variance, maximum, minimum, kurtosis, and skewness, to capture the basic statistical characteristics of the data; S402: Extract trend and seasonal components of time series using moving average method; S403: Perform Fourier transform on the time series data to obtain a spectrum, analyze the frequency components, and extract the dominant frequency and energy distribution characteristics in the spectrum; S404: Extract key features from the time-frequency spectrum using short-time Fourier transform method; S405: Use the chi-square test method to evaluate the correlation between the features and the target variable, select the features with higher correlation, and remove irrelevant and redundant features; S406: Using the extracted and selected features as input variables, and using the state and key parameters of the energy storage system as target variables; S407: Divide the data set into a training set and a test set; S408: Initialize the random forest model and set parameters including the number of trees, maximum depth, and minimum number of sample splits; S409: Use the training set data to train the random forest model and construct multiple decision trees. Each tree randomly selects features for splitting during training. S410: Use the trained random forest model to predict the test set data, calculate the accuracy, recall rate, and F1 score indicators of the prediction results, and evaluate the performance of the model; S411: Analyze the importance of each feature in the random forest model, identify the features that have a greater impact on the state or key parameters of the energy storage system, and identify potential problems of the energy storage system based on the importance of the features.

5. The method for optimizing the charging and discharging strategy of an energy storage system based on big data analysis according to claim 4 is characterized in that: Based on the data analysis results, a time series prediction model and a particle swarm algorithm optimization model are established; the specific steps are as follows: S501: According to the data processed in step S3, a time series to be predicted is selected to build a time series prediction model; S502: Construct the autoregressive part of the prediction model. The formula is as follows: X t =φ1X t-1 +φ2X t-2 +...+φ p X t-p +∈ t ; Among them, X t is the value of the time series at time t, φ i is the autoregressive coefficient, p is the autoregressive order, ∈ t It is white noise; S503: Construct the moving average part of the prediction model. The formula is as follows: ∈ t =θ1∈ t-1 +θ2∈ t-2 +...+θ q ∈ t-q ; Among them, θ i is the moving average coefficient, q is the moving average order; S504: combining the autoregressive and moving average parts to form an ARIMA (p, d, q) model, where d is the difference order; S505: Use time series data to estimate the parameters of the ARIMA model, including autoregressive coefficients, moving average coefficients, and difference orders, and use the least squares method to optimize the parameters; S506: According to the trained ARIMA model, the future time series data is predicted, and the accuracy of the prediction results is evaluated using the mean square error; S507: randomly generate a group of particles, each particle represents a potential charging and discharging strategy solution, including key parameters of charging time, discharging time, and charging and discharging power; S508: Initialize the speed and position of each particle, where the speed represents the moving direction and size of the particle in the search space, and the position represents the current solution of the particle; S509: Use the time series prediction model to predict the state of the energy storage system in the future; S510: Calculate the fitness value of each particle according to the prediction result; S511: Update the speed and position of each particle according to the fitness value of the particle and the global optimal solution, where the speed update formula is as follows: in, is the velocity of particle i at the k+1th iteration, ω is the inertia weight, c1 and c2 are learning factors, r1 and r2 are random numbers, is the personal optimal solution of particle i, g k Make the global optimal solution, is the current position of particle i; The position update formula is as follows: S512: Repeat S509-S511 until the upper limit of the number of iterations is reached or the fitness value meets the preset conditions, and output the global optimal solution.

6. The method for optimizing the charging and discharging strategy of an energy storage system based on big data analysis according to claim 5 is characterized in that: According to the results of the prediction model and optimization model, formulate a preliminary charging and discharging strategy, including the key parameters of charging time, discharging time, and charging and discharging power; the specific steps are as follows: S601: Obtain the energy storage system status prediction for a period of time in the future from the time series prediction model, including but not limited to the state of charge, temperature, and power demand of the battery pack; identify the peak demand, valley period, and possible abnormal conditions that may occur in the future, and determine the feasible range and safety threshold of charging and discharging according to the characteristics and limitations of the energy storage system; S602: Setting initial charging and discharging strategy parameters, including charging time, discharging time, and charging and discharging power; S603: Optimizing the charge and discharge strategy parameters based on the prediction data using a particle swarm optimization algorithm; S604: Continuously adjust the charging and discharging strategy parameters through iterative calculation until an optimal solution is found or a preset upper limit of iteration times is reached, and output the result; S605: Formulate a detailed charging and discharging plan according to the optimized charging and discharging strategy parameters, including specific charging time, discharging time, and charging and discharging power.

7. The method for optimizing the charging and discharging strategy of an energy storage system based on big data analysis according to claim 6 is characterized in that: Use historical data and real-time data to simulate and verify the strategy, evaluate the performance and effect of the strategy, and adjust and optimize the strategy based on the simulation verification results; the specific steps are as follows: S701: Build a simulation platform to simulate the operation of the energy storage system, including the charging and discharging behavior of the battery pack, the working status of the PCS, and the monitoring function of the BMS; S702: Importing historical data and real-time data into the simulation platform; S703: according to the established charging and discharging strategy, setting simulation parameters, including charging time, discharging time, and charging and discharging power; S704: Setting the start and end time of the simulation, as well as the time step of the simulation; S705: Starting the simulation platform to simulate the operation of the energy storage system according to the set parameters and strategies; S706: Real-time recording of state variables of the energy storage system during the simulation process, including the state of charge, temperature, voltage, current, etc. of the battery pack, and the input and output power and efficiency of the PCS; S707: Compare the simulation results with the expected goals, evaluate the performance of the charging and discharging strategy, and calculate key performance indicators, including the energy utilization rate, charging and discharging efficiency, temperature control effect, and power response speed of the energy storage system; S708: Analyze abnormal situations or unexpected behaviors that occur during the simulation process and identify possible problem points; S709: adjusting and optimizing the charge and discharge strategy according to the simulation evaluation results, including adjusting key parameters of the charge time, discharge time, and charge and discharge power; S710: Repeat the simulation until the performance of the strategy reaches the expected target or cannot be further improved; S711: Output the optimized charging and discharging strategy as executable instructions and schedules for use by the energy storage system controller.

8. The method for optimizing the charging and discharging strategy of an energy storage system based on big data analysis according to claim 7 is characterized in that: Use big data technology to establish a fault prediction model to monitor and predict key components of the energy storage system in real time, and promptly detect and handle potential faults. The specific steps are as follows: S801: Use random forest model to build fault prediction model; S802: Use the training set divided by historical data to train the model, use the trained random forest model to predict the test set data, and adjust the model parameters and feature selection according to the performance of the model on the test set to optimize the performance of the model; S803: Deploy the trained fault prediction model to the monitoring platform of the energy storage system to monitor and predict key components in real time; S804: Deploy sensors and monitoring equipment on key components of the energy storage system to collect real-time operating data, including voltage, current, temperature, pressure, vibration parameters, as well as working status and mode operation information; S805: The collected data is transmitted to the data analysis platform in real time via wireless means; S806: Input the real-time collected data into the random forest fault prediction model to perform real-time fault prediction. The model will predict whether there is a risk of failure of key components based on the changes and trends of the data; S807: When the model predicts a failure risk, the system sends a warning signal to notify the operator.

9. The method for optimizing the charging and discharging strategy of an energy storage system based on big data analysis according to claim 8, characterized in that: The construction of the random forest fault prediction model includes the following steps: S901: For each decision tree, randomly select a portion of features for splitting, and calculate the splitting point of each feature; S902: Divide the data set into different subsets according to the splitting points, and recursively build a subtree on each subset to obtain a complete decision tree; S903: In the prediction phase, new data is input into the random forest model, and each decision tree will give a prediction result; S904: adopting a voting mechanism to select the prediction result of the majority decision tree as the final prediction result.

10. The method for optimizing the charging and discharging strategy of an energy storage system based on big data analysis according to claim 9, characterized in that: Monitor the energy storage system in real time, collect operating data and feed it back to the data analysis platform, and regularly evaluate and adjust the charging and discharging strategy; the specific steps are as follows: S1001: Sensors and monitoring equipment collect operating data of the energy storage system in real time; S1002: Perform preliminary data cleaning at the data collection end to remove obviously erroneous or invalid data points, and transmit the pre-processed data to the data analysis platform in real time; S1003: On the data analysis platform, the energy storage system operation data collected in real time is stored in a database; S1004: Use data visualization tools to display the operating status and key parameters of the energy storage system in real time: S1005: setting an evaluation cycle of a charging and discharging strategy according to the operating characteristics and actual needs of the energy storage system; S1006: At the end of the evaluation cycle, collect and organize the operation data of the energy storage system during the cycle, including real-time collected data, historical data, and simulation data; S1007: Analyze the implementation effect of the charging and discharging strategy during the cycle using big data analysis, including key performance indicators of the energy storage system, such as energy utilization rate, charging and discharging efficiency, temperature control effect, and power response speed; S1008: Compare the system performance changes before and after the implementation of the strategy, evaluate the effectiveness of the strategy, and identify problems in the operation of the energy storage system based on the strategy effect analysis; S1009: According to the problem identification result, the charging and discharging strategy is adjusted and optimized accordingly, including adjusting key parameters of charging time, discharging time, and charging and discharging power; S1010: Use the simulation platform to verify the adjusted strategy and evaluate its performance and effect; S1011: Update the verified optimized charging and discharging strategy to the control system of the energy storage system.

Images